I agree that the oracle problem makes many use cases for Ethereum over-hyped at present. Nevertheless, there are plenty of oracles that can be easily shaken out, such as the Fed's federal funds target rate, or the inflation rate as reported by the Bureau of Labor Statistics, or simply an asset price from an exchange or group of exchanges. There is a business for this, and http://www.oraclize.it/ is probably going to do well. There will be more.

But that's not really the short- or mid-term promise of Ethereum, IMHO. Ether is programmable money, and within the context of the transfer of value via ETH, payment networks and mechanisms can be highly granular and complex because of the trustless and open nature of the ledger, and because the Ethereum scripting language is Turing-complete. As an example, Bitcoin Core developers are hoping to create payment channels for Bitcoin (called the Lightning Network), which will require many changes to the Bitcoin code, as well as at least one hardfork. Such payment channels can be added to Ethereum as is. In addition to being Turing-complete, Ethereum uses an account-based architecture for accounts balances rather than the UTXO (unspent transaction output) model that Bitcoin employs. The balance at an ether address is expressed in a constant number of bytes due to settled deposits and withdrawals, Bitcoin uses the sum of all unspent transaction outputs to determine the balance of an address, which can create problems for bundling lots of small payments on-chain as some UTXO transactions can be big. Also, Ethereum blocks are 14 seconds with little variance, whereas Bitcoin blocks are 10 minutes with high variance. In short, in just about every use case, Ethereum is better for transacting ETH than bitcoin is for transacting BTC.

Also, you can simply use the Ethereum network as a ledger without intrinsic contract triggers that need oracles, like tracking the ownership of shares.

EDIT: I went over to HN and dropped a reply. Jesus, those HN folk hate cryptocurrency. They have been powerful haters of Bitcoin all along, and it looks like to be the same case for Ethereum. It's very interesting, because you would think that a forum of programmers would be a generally receptive audience. There are a lot of misinformed and negative comments every time cryptocurrencies are discussed.